Sediment Classification Using Image Processing
Sediment classification is required for identifying type of sediment present near the wall of dam. As the flow of
water through gates of dam is irregular, sediments flow out through them causing damage to gates. In order to avoid this we
are classifying the sediments depending on their sizes and accordingly manage the flow through gate. Presently sediments
are classify using sonar imaging system, this system will transmit sound signal from surface of water and echoes are
recorded. These echoes are used to plot the image of sub bottom profile of oceans. These systems are very costly and takes
larger time for processing. The proposed system is using image of sediments. In this paper it is proposed that sediment
classification is done into three different categories as rocky sediments which is having size greater than 25 mm, gravels that
is medium size sediment having size between 2-8 mm. Coarse sand having grain size less than 2 mm. Input to the system
will be the image of sediment. This input is given to system and it will extract the feature of input sample and it will try to
match with database using artificial neural network. We are proposing an adaptive algorithm in which only the visual
properties of sediments is used for classification. This will reduce the time for processing and make it easier for the adaptive
algorithm to learn. The results are obtained on the actual data. Software Requirement for this system is MATLAB 2013b.
The output of system is classification of sediments, which is much faster and accurate than present system.